Exploring Topographic Effects on Surface Parameters over Rugged Terrains at Various Spatial Scales
Hanyu Shi,Zhiqiang Xiao
DOI: https://doi.org/10.1109/tgrs.2021.3098607
IF: 8.2
2021-01-01
IEEE Transactions on Geoscience and Remote Sensing
Abstract:Topography is an inevitable factor when processing remote sensing data. Slope and aspect are sufficient for describing topographic conditions within a fine-scale pixel (e.g., 30 m); the resulting schematic is referred to as a sloping terrain and is modeled as a solo slope. A composite slope, which contains many solo slopes that are collectively referred to as rugged terrain, is needed for coarse-scale pixels (e.g., 1 km). However, many parameter estimation algorithms use topographic approximation methods, such as the assumption of a flat surface, assumption of a solo slope, omission of contributions from adjacent slopes, and usage of the terrain view factor (TVF) to approximate adjacent contributions. These topographic approximations can induce significant errors over mountain areas; however, errors caused by various approximation methods have not been comprehensively analyzed. This study summarizes radiative transfer (RT) processes over rugged terrains, proposes composite-slope models for surface parameters, and analyzes the influences of different topographic approximation methods on surface reflectance ( ${\rho }$ ), directional brightness temperature ( ${T_{b}}$ ), surface net radiation ( ${E_{n}}$ ), slope downward radiation ( ${E_{d}}$ ), absorbed photosynthetically active radiation (APAR), total emitted solar-induced chlorophyll fluorescence (SIF) by all leaves ( ${F_{e}}$ ), SIF observed at the top of the canopy ( ${F_{o}}$ ), broadband albedo ( ${\alpha }$ ), and broadband hemispherical emissivity ( ${\varepsilon }$ ) at a series of spatial resolutions (30, 90, 270, 540, 1080, and 5400 m). Three surface types are tested: vegetation, soil, and snow. The results demonstrate that: 1) assumptions of a flat surface or a solo slope and the use of the TVF method induce significant errors (1%–58%) in all aforementioned parameters; 2) adjacent contributions can be neglected when simulating ${\varepsilon }$ , APAR, ${F_{o}}$ , ${F_{e}}$ , and low-reflective ${\rho }$ ; and 3) adjacent contributions should be considered for ${E_{n}}$ , ${E_{d}}$ , and high-reflective ${\rho }$ , and they are also significant when simulating ${\alpha }$ using fine-resolution data or over snow surfaces. These findings and the composite-slope models developed in this study benefit those who intend to conduct forward modeling and parameter estimation studies over rugged terrains.